Questions tagged [neural-networks]
Questions involving defining, training, executing, importing and exporting neural networks.
593 questions
3
votes
1
answer
74
views
Does NetTrain automatically shuffle the training data you input into it?
I'm trying to train a network on unshuffled data in a specific order (I know this is really bad for accuracy, it's for an experiment). Does anyone know if NetTrain shuffles training data automatically ...
- 31
3
votes
1
answer
130
views
NetTrain: One or more inputs provided to port "Output" was invalid: input is not an integer between 1 and 2
Thank you for seeing this question. I was trying to train a LSTM network when the error occurred:
NetTrain::skipbatch: Batch #1 will be skipped, because one or more inputs provided to port "...
1
vote
0
answers
66
views
Error in Scoring net for an AttentionLayer
I am unable to determine where this error is coming from:
AttentionLayer::netmissio: Net specified for the scoring net (argument 1) is missing required input port Query.
...
- 425
0
votes
0
answers
81
views
How to tailor the input of a predefined network to match my own data?
I'd like to use the ResNet V2 on my own data, but my data is not image. Let's say my data is a list of numbers or a time sequence. Therefore, I decided to revise the ResNet V2 a little bit.
...
6
votes
5
answers
339
views
Data conversion
I have data in a text file in the form of:
{5,15,23,28,36}->{2,8,10,14,27}
I would like to convert it to this form
...
- 425
3
votes
0
answers
349
views
Neural Networks for stock prediction
I'm working on a binary classification model in Mathematica to predict whether the price of a stock will go up or down in the next 30 candles (1 hour ahead, using 2-minute intervals). I'm using LSTM ...
- 3,146
2
votes
0
answers
95
views
Global change of embedding vector length in Wolfram models
When using one of the models from the Wolfram net repository, is there an easy way to change the length of the embedding vector throughout the model in order to train the model using an alternate ...
- 415
4
votes
1
answer
335
views
Untrained neural network models
Wolfram says they make available untrained neural net models, but as far as I can tell, all the models in the neural net repository are trained. Where are the untrained models?
- 415
0
votes
1
answer
157
views
Shared Weights in Neural Nets
The output from a neural net wants to be W.Transpose[W]+b. Back-to-back Linear one without bias is close, but the weight matrices will be different. Is there a way of accomplishing the desired net?
- 415
1
vote
0
answers
59
views
Is this a reasonable way to transform simple MSE loss function to weighted MSE for neural networks?
The idea is tested on a dataset with a clear outlier (see picture below)
...
- 524
0
votes
0
answers
93
views
Combination of decoder and encoder for VAE interpolation
To get VAE interpolation I need decoder and encoder to be combined. Can you make a suggestion?
...
- 11
0
votes
1
answer
83
views
NetTrain fails with MXNetError: Check failed: assign(&dattr, vec.at(i)) on a complex residual network (CNN)
I am designing a relatively large CNN using many convolution layers with resedual blocks. My attempt to train this network always fails with
...
- 1,769
0
votes
1
answer
455
views
Implementation of PINNs (Physics Informed Neural Networks) in Mathematica
I know Mathematica does not have an Autograd function for differentiating a neural network with respect to its input. Has any one tried an implementation PINNs for solving PDEs in mathematics?
Many ...
- 21
0
votes
1
answer
114
views
Scheduled Updating of LossFunction Scaling in NetTrain
I'm looking for a way to externally manipulate (hyper)parameters of a network during NetTrain training, as a function of the training progress. Specifically, I want ...
- 347
1
vote
0
answers
75
views
Is there any way to run segmentation model SAM2?
I would like to run sam2: https://github.com/facebookresearch/segment-anything-2.
Does ServiceConnect allow us to connect to meta.ai or llama 3.1 to do things like ...
- 3,820